Blackman&Robalino.Adoption Of Soil Conservation

Blackman&Robalino.Adoption Of Soil Conservation

• 1.
  Adoption of soilconservation practices by Costa Rican coffee farmers Allen Blackman and Juan Robalino EfD Center for Central America 2 nd Annual Environment for Development meeting Beijing, November 3-7, 2008
• 2.
  Policy context Academiccontext Data (Very) preliminary econometric results Simple probits Spatial regressions Looking ahead Outline
• 3.
  Coffee is aleading ag. commodity in Costa Rica 57,000 growers 100,000 hectares $ US 200 million/year export revenues annually Serious environmental impacts Most Latin American coffee is shade grown prevents soil erosion harbors biodiversity sequesters carbon facilitates aquifer recharge But 90% Costa Rican coffee is “technified” Minimal shade, agrochemical intensive, higher yields Environmental impacts Soil erosion Sedimentation and eutrophication of streams Contamination surface and groundwater Policy context
• 4.
  Soil conservation practices(adoption rates Turrialba/C.B.) contour planting (19%) infiltration holes (12%) deviation canals (23%) vegetative barriers (7%) Promoting further adoption challenging significant fixed costs long payoff periods 90% of growers small-scale growers (<100 quintals/yr) What drives adoption? grower characteristics: e.g. age, education? farm characteristics: e.g. size, slope, climate? spatial spillovers: e.g. learning? Policy context (cont’d)
• 5.
  Extensive econometric literatureson adoption of soil conservation practices spatial spillovers in technology adoption Little or no econometric literature spatial spillovers in adoption of conservation practices adoption of conservation practices by coffee farmers Meta-analysis of 31 studies of adoption of conservation practices (Knowler & Bradshaw 2007) only a handful include variables aimed at capturing spillovers, and none use spatial econometric approach none on coffee Academic context
• 6.
  First source: NationalStatistics and Census Institute (INEC) and CR Coffee Institute (ICAFE) census of coffee growers coverage: Turrialba & Coto Brus (2003) n ~ 6,200 dependent variables contour planting infiltration holes deviation canals vegetative barriers independent variables grower characteristics (sex, age, education, etc.) farm characteristics (size, variety, non-coffee ag., etc.) geophysical characteristics (temperature, precipitation) location (boundaries and centroids) Data
• 7.
  Second source: GISfrom a variety of sources elevation aspect (directional orientation) average slope Holdridge life zones distances to coffee markets population centers neighbors Data (cont’d)
• 8.
  Probit regression results
• 9.
  Spatial autocorrelation: contour planting adoption dummy
• 10.
  Spatial autocorrelation: Errors from OLS regression
• 11.
  Spatial autocorrelation: Errors from probit regression
• 12.
  Probit, OLS, Anselin& LeSage regression results
• 13.
  Probit and IVregression results
• 14.
  Thank you